List your company

Question Answering on MMLU English Large Shift (test)

44.4AccuracyTemp
36.39238.47140.5542.629Jan 28, 2026

Evaluation results

MethodLinks
Temp2026.0144.416.11.32
BLoBEnsemble size (N)=102026.0142.322.31.81
MAP2026.0141.238.42.19
Dropout2026.0139.638.62.14
LA2026.0139.112.81.27
LLLA2026.0138.939.71.27
Ckpt Ens2026.0138.732.51.72
Bayesian-LoRASample size (S)=4, Training protocol=End-to-End2026.0137.520.41.26
BBB2026.0136.722.41.83